EXPERIMENTS RELATIVE TO METEORITES. 337 



finish tliem from all other silicated rocks, and which deserve to fix the attention. 

 (1) l\'ridut represents the most basic silicated type that is known, whether in 

 the meteorites or in eruptive rocks. In that series of which it constitutes the 

 first term and which terminates at the granite, it forms the species at once the 

 most simple in composition and the best defined. (2) Under the point of view 

 of the mode of crystallization, })cridot, as well as the bisilicate of magnesium, or 

 estatite, which is its frequent companion, is distinguished from the aluminous 

 silicates, particularly^ from those of the group i)f feldspar, by the facility with 

 which they are formed and become crystallized by the dry way, after simple 

 fusion. On the contrar}*, it has never been foimd practicable to crj^stallize arti- 

 ficially, under the same conditions, anything having a resemblance, even i-emotely, 

 to feldspar and to granite. (3) Hocks of jjcridot are very remarkable also by 

 their great density, which is superior, as the following table will show, to that 

 of all other eruptive rocks, and even to that of the basalts : 



Granite 2.64 to 2.76 



Trachyte 2.62 to 2.88 



Porphyrite 2.76 



Diabase 2.66 to 2.88 



Basalt 2.9 to 3.1 



Enstatite 3.303 



Lherzolite 3.25 to 3.33 



Peridot 3.33 to 3.35 



These diflFerent rocks must at their origin have been superposed on one another 

 in an order conformable to their augmentation of density. The great density of 

 the peridotic rocks justifies the normal position wliich they appear to hold within 

 the crust of the earth, beneath the granite envelope, beneath even the aluminous 

 basic rocks. 



COMPAKATIVE DENSITIES OF METEORITES AKD OF THE PRINCIPAL TER- 

 RESTRIAL ROCKS. 



Setting aside the carbonaceous meteorites, which should be considered as out- 

 side of the series, we might conceive the meteorites to be disposed in concentrio 

 spherical strata, forming an ideal globe, whose density would go on increasing 

 from the surface towards the center. At the exterior would be the aluminous 

 stones ; then would come the jieridotic stones, those of the common type, tho 

 polysiderous, the syssiderous, and finally the holosiderous. 



AVe may remark that this theoretical section is not without some analogy with 

 an ideal section of the terrestrial globe, distinction being made of the sedimen- 

 tary deposits and of thegranito-gneissic stratum. In this section the lavas would 

 correspond to the aluminous meteorites ; l)elovv these the peridot would be tlio ^ 

 analogue of the stone of Chassigny ; the lherzolite and other rocks of the same 

 kind would well repieseut the meteorites of the connnon type. Here, it is true, 

 stop the analogies which can be directly observed, but here also stops the knowl- 

 edge which we have of the more profound regions of our globe. There is 

 nothing repugnant to reason in supposing that these more profound parts of tho 

 earth afford resemblances to the ideal globe which we have just constructed by 

 the superposition of tho diflPerent kinds of meteorites ; there is nothing to prove, 

 in a word, that one of the globes does not complete the other. 



This comparison, which may seem, perhaps, somewhat precarious, will be bet- 

 ter understood by means of the following table, the first colunm of which con- 



22 s 



